Abstract B028: Assessing variability in training deep models in delineating Prostate gland anatomy between clinical experts
Bibliographic record
Abstract
Abstract Background: Deep learning (DL) methods provide enormous promise in automating manually intense tasks such as medical image segmentation and provide decision support in clinical workflow. The deep neural network requires a large amount of training examples and a variety of expert opinions to build a reproducible model. The problem becomes challenging to train these deep models with differing expert provided opinions. Inter-reader variability among clinical experts is a real-world problem that severely impacts the generalization of deep networks. Methods: This study proposes quantifying the variability in deep network performance using expert opinions and exploring strategies to train the network and to adapt between experts. We propose to address the inter-reader problem in the context of prostate gland segmentation using a well-studied deep 3D U-Net architecture. Two expert reader groups (R#1, n=342 and R#2, n=204) provided the clinical annotation that formed the reference data, where the prostate glandular anatomy was delineated in a Magnetic resonance imaging (MRI, T2-Weighted). The experts groups were from two-independent clinical centers provided annotation on the same patients imaging, with the members of the two teams blinded on the others opinion. Results: The network was trained and tested with individual expert examples (R#1 and R#2) and had an average dice coefficient of 0.825 (CI: [0.81 0.84]) and 0.85(CI: [0.82 0.88]), respectively. Combined training with a representative cohort proportion (R#1, n=100 & R#2, n=150) provided improved model reproducibility between readers in the test cohort with an average dice coefficient of 0.863(CI: [0.85 0.87]) (for R#1) and 0.869(CI: [0.87 0.88]) (for R#2). We find the model performance improved in a subcohort with large gland volumes; the best dice for R#1 and R#2 were 0.846 [CI: 0.82 0.87] and 0.872 [CI: 0.86 0.89], respectively estimated using 5-fold cross-validation. Conclusion: We find proportional representation of multiple annotators improved the performance across readers and help generalization of the networks. Citation Format: Shatha Abudalou, Yasin Yilmaz, Jung Choi, Yoganand Balagurunathan. Assessing variability in training deep models in delineating Prostate gland anatomy between clinical experts [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Artificial Intelligence and Machine Learning; 2025 Jul 10-12; Montreal, QC, Canada. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(13_Suppl):Abstract nr B028.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.024 | 0.084 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".